Skip to content
All library documents

On-the-Run and Off-the-Run Bond Yield Spreads

Article Quant Q&A · Author: low_snr

Summary

The document considers whether the yield difference between on-the-run and off-the-run bonds of similar maturity could support a mean-reversion trade. It explains why the spread may reflect persistent differences in liquidity and investor demand: new issues tend to be more liquid, index-tracking investors may need to buy them, and liquidity-focused trading may concentrate in them. Buybacks and bond splitting can also affect relative demand.

The bonds can have different coupons because they were issued under different interest-rate conditions. That means bonds with similar maturities can still differ in duration and convexity, making their relative yields respond differently as rates change. A spread trade may therefore face changing risk exposures and margin pressure when held with leverage. The response describes convergence as potentially very slow, possibly requiring a long holding period; it offers qualitative considerations rather than data testing the spread's stationarity or trade profitability.

Key ideas

  • On-the-run bonds often have a liquidity advantage over comparable off-the-run issues.
  • Index demand and liquidity-seeking traders can affect the relative yields of the two bonds.
  • Different coupons create duration and convexity differences even when maturities are similar.
  • Rate moves can change the spread and create margin pressure for leveraged positions.
  • Convergence may take a long time, so the spread is not established as a short-horizon mean-reversion opportunity.

Tags

Full text
# ONTR and OFFR of Similar Maturity


# ONTR and OFFR of Similar Maturity












I was wondering if anyone has experience trading a mean-reversion strategy between on-the-run bond and off-the-run bonds of similar maturity.

My expectation is that these two bonds would move in sync, implying that the difference in their yields is stationary. However, since OFFR bond yield data is not readily available, I haven’t been able to test this hypothesis. I’d appreciate any insights on the following:

- Is it correct to assume that the difference in yields between ONTR and OFFR bonds is stationary, with flows being the primary factor affecting the yield difference?

- What idiosyncrasies could cause the spread to exhibit momentum and potentially break the stationarity of the time series? For instance, one scenario I considered is if one of the bonds becomes part of the deliverable basket for a bond future, causing increased demand and a sharp drop in its yield. Are there any other factors that could explain such behavior?

- What do you think about the feasibility of such a mean-reversion strategy? What are the main challenges?

## Answer by AlRacoon (score 2, accepted)

https://quant.stackexchange.com/a/81671

As both you and the commenter identify, there are market dynamics that could cause one bond to be more or less preferable than the other. In general, ONTR bonds, as new issue, tend to be more liquid then OFFR bonds. One of the primary reasons for this is that many "real money" investors are managing assets to a "market value weighted" index. As such, the newly issued bonds need to be bought by such investors to match the index weights and reduce their tracking error to their benchmark. This activity tends to put pressure on OFFR bonds and increase demand for ONTR bonds. Also, as many "fast money" (aka hedge funds) strategies require liquidity, they tend to be manifested in the ONTR bonds. There are other market dynamics that could affect the liquidity as well, such as buybacks, splitting activity to meet insurance company needs etc.

With respect to bond risk, OFFR bonds will tend to have different risk than ONTR bonds with the same maturity. The OFFR were issued in a different interest rate environment than the current ONTR bonds. As these bonds are auctioned at par, they are going to have yields reflective of the market at time of auction. Said another way, they will tend to have different coupons depending on when they were issued. As such, they will have different duration and convexity characteristics than the ONTR bonds of identical maturity. This will cause the spread between these bonds to change depending on what happens to interest rates. And consequently potential margin activity if one were to play this spread convergence on a levered basis.

As a practical matter, one would have to be a long term investor to be able to hold these bonds until complete convergence. ONTR notes/bonds are issued at set maturities. To get convergence, one may have to wait as long as 20Yrs to get complete convergence at maturity (assuming one is playing this strategy using formerly 30Yr bonds that have rolled down to 20Yr maturity ONTR bonds).

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.